On 12/14/20 10:05 AM, Bruno Haible wrote:
But canonicalize-lgpl.c had similar changes, and it has a similar failure
on AIX 7.2, still today:
../../gltests/test-canonicalize-lgpl.c:100: assertion 'result == NULL' failed
FAIL test-canonicalize-lgpl (exit status: 134)
Probably your changes
Paul Eggert wrote:
> > This patch introduced a unit test regression on AIX 7.2
> > ../../gltests/test-canonicalize.c:112: assertion 'result1 == NULL' failed
> > FAIL test-canonicalize (exit status: 134)
> ...
> I installed the attached patches to work around these issues, along with
> other st
On 12/11/20 3:47 AM, Bruno Haible wrote:
This patch introduced a unit test regression on AIX 7.2
../../gltests/test-canonicalize.c:112: assertion 'result1 == NULL' failed
FAIL test-canonicalize (exit status: 134)
Thanks for reporting it. My recent changes to the canonicalize code assumed
Hi Paul,
> 2020-12-02 Paul Eggert
>
> + canonicalize: fix EOVERFLOW bug
> + * lib/canonicalize.c (canonicalize_filename_mode):
> + When testing whether a directory entry is a symbolic link, or a
> + directory or other, do not use lstat or stat or
> + areadlink_with_size.
* lib/canonicalize.c (canonicalize_filename_mode):
When testing whether a directory entry is a symbolic link, or a
directory or other, do not use lstat or stat or
areadlink_with_size. Just use areadlink, as this suffices and it
avoids the EOVERFLOW problem that lstat and stat have.
* modules/canon
